Dual-axis image equalization for video conferencing perspective correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video conferencing systems face challenges in maintaining uniform participant size and perspective due to depth of field disparities, making it difficult for participants seated far away from the camera to be clearly visible and recognizable.
Innovation Solution
A video conferencing system employing dual-axis image equalization using a processor that receives images from two cameras positioned at determined angles, applies dual-axis warping functions to each image independently, and stitches them together to create a display image that corrects perspective and size discrepancies, ensuring all participants appear similar in size and proportion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a single camera captures participants at different distances from the camera, then the camera can capture the entire scene, but participants at different distances appear at different sizes due to depth of field disparities
Solution Approach 1:
The patent divides the scene into multiple sub-scenes captured by multiple cameras positioned at different locations. Each camera captures a specific portion of the scene, and the system processes each sub-scene independently before combining them. This segmentation allows each camera to focus on a limited area, reducing the depth of field disparity within each sub-scene while maintaining comprehensive coverage of the entire conference room.
Solution Approach 2:
The patent applies different processing parameters to different regions of the image. Each sub-scene captured by a specific camera is processed with warping functions tailored to that camera's position and field of view. This local quality approach ensures that each region is optimized for its specific characteristics, allowing participants in different locations to appear at uniform sizes in the final composite image.
2Area of stationary object
If the camera uses a wide field of view to capture all participants, then all participants are visible, but perspective distortions increase making participants at different distances appear at different sizes
Solution Approach 1:
Instead of using a single wide-angle camera that captures the entire scene and suffers from severe perspective distortions, the patent segments the scene into multiple smaller fields of view captured by multiple cameras. Each camera captures a narrower portion of the scene, minimizing perspective distortion within each sub-scene. The system then combines these less-distorted sub-scenes to create the final image, achieving both wide coverage and reduced distortion.
Solution Approach 2:
The patent transitions from a single-camera wide-angle view to a multi-camera system that captures the scene from multiple spatial positions. This dimensional change from one viewpoint to multiple viewpoints allows the system to reconstruct the scene with reduced perspective distortion, as each camera provides a less distorted view of its local area.
3Shape
If multiple cameras are used to capture different portions of the scene, then perspective distortions are reduced, but the system complexity increases
Solution Approach 1:
The patent merges multiple camera inputs into a single unified output image. The system captures sub-scenes from multiple cameras, processes each sub-scene independently to correct perspective distortions, and then combines the processed sub-scenes into a final composite image. This merging approach maintains the benefits of multiple cameras (reduced distortion) while presenting a simplified output that appears to be captured from a single viewpoint.
Solution Approach 2:
The patent introduces an image processing system as an intermediary between the multiple cameras and the final output. This intermediary processes each sub-scene from the multiple cameras, applying warping functions and combining them into a unified image. The intermediary handles the complexity of managing multiple camera inputs and integrating them into a coherent output, simplifying the overall system architecture.
Data Source
AI summary
A video conferencing system is provided that reduces the perspective of an image captured by a video conferencing camera. The perspective is reduced in an image by receiving a first image of a first field of view and receiving a second image of a second field of view. Moreover, the first image is equalized and the second image is equalized independently of the first image. Additionally, the first equalized image is stitched to the second equalized image to create a display image, which can be sent to a display another video conferencing endpoint.